magician

Systematic performance profiling and optimization with a mandatory baseline-first gate — measure before changing, re-measure after. Use when something is slow or you need to hit a latency/throughput target.

Blameless post-mortem / RCA — gathers facts, reconstructs a timeline, runs 5-Whys, defines action items, writes the post-mortem file and commits it. Use after an incident or outage.

Structured design dialogue with a visual companion — produces an approved spec and design artifacts before any implementation begins. Use at the start of a feature, before writing code.

Thorough, research-grounded code review of a change, PR, or MR — multi-lens (correctness, security, simplification, tests), severity-ranked with impact + fix, configurable depth, optional PR comments. Use when asked to review code, "do a code review", "review this PR/MR", "review my changes/diff/branch", or audit a changeset before merge.

Use when the user asks to research, investigate, analyze, find out, explore, examine, audit, or evaluate something — structured multi-source research with consulting, library-doc search, web search, and guided output delivery.

Multi-agent code review AND remediation — dispatches correctness, security, and simplification reviewers in parallel, consolidates findings, then fixes criticals/highs. Use when reviewing a diff or PR before shipping.

Detect a repository's languages, frameworks, databases, and observability stack, then load only the relevant Magician lore cores and task-matched deep dives. Use before implementation, debugging, review, performance, database, or logging work when project-specific technical guidance would improve the result, or when the user asks to inspect the stack, load lore, or establish project context.

Enable, configure, or disable the Magician CLI status line ("Magician Claude CLI UI") — a lightweight, always-on bar showing live context %, a context-rot warning, a token-flow sparkline, model · git · cost, the active skill/workflow/loop, the reasoning effort/mode, the bundled-lore state, and the output-brevity "voice" level. Also sets the output-brevity voice (warrior/scribe/bard). Use when the user says "enable/turn on the magician UI / status line / status bar", "show my context/tokens/effort in the console", "configure/change what the bar shows", "what should the status line display", "turn off / disable the status line", or "make output shorter/leaner / set the voice / reduce token cost".

Systematic performance profiling and optimization with a mandatory baseline-first gate — measure before changing, re-measure after. Use when something is slow or you need to hit a latency/throughput target.

One-time Codex workspace setup — creates .workspace/ structure, .gitignore entries, a lean AGENTS.md, and suggests relevant Codex capabilities. Run once per project.

Blameless post-mortem / RCA — gathers facts, reconstructs a timeline, runs 5-Whys, defines action items, and writes a post-mortem for review. Use after an incident or outage.

Converts an approved spec into a TDD task plan with a parallelism map (PARALLEL vs SEQUENTIAL), saved to .workspace/shared/plans/. Use after a spec is approved, before implementation.

Full verification loop — tests, types, lint, build, and a Playwright browser check for UI projects; collects evidence before any success claim. Use to verify a change is actually green.

Memory & context steward — session-learning history, the global reference store (repos/projects/ideas), AND live context management (size status, post-compaction resume capsule, project learnings, promotion). Use to review past sessions, remember/recall/forget a reference, check context size, resume after compaction, or capture/consolidate learnings.

Work with Confluence over its REST API (no MCP) — read/search (CQL), summarize pages, find docs, create/update/comment/label. Use for any read/search/create/update on Confluence pages, or references to a remembered space, page, or doc.

Structured design dialogue with a visual companion — produces an approved spec and design artifacts before any implementation begins. Use at the start of a feature, before writing code.

Gated CI/CD pipeline management for GitHub Actions and GitLab CI, plus configuration review for other providers when their authenticated tooling is available. Use to set up or fix CI/CD.

Thorough, research-grounded code review of a change, PR, or MR — multi-lens (correctness, security, simplification, tests), severity-ranked with impact + fix, configurable depth, optional PR comments. Use when asked to review code, "do a code review", "review this PR/MR", "review my changes/diff/branch", or audit a changeset before merge.

Creates a new reusable skill — can be suggested by the pattern detector after repeated requests. Use to scaffold a new SKILL.md.

Work with Jira over its REST API (no MCP) — read/search (JQL), look up tickets, my board/sprint, create/comment/@mention/transition/link/worklog, MR investigation, clone the ticket's repo. Use for any read/search/create/update/transition on Jira issues, or references to a remembered board, project, epic, or person.

Local code knowledge-graph + cache (no MCP, no network) — index a repo and retrieve ranked file:line for a topic instead of grepping whole files; show graph/cache status; blast-radius / dependents of a file; refresh or reset. Use for "knowledge graph status", "index this repo", "what depends on <file>", "find the code for <thing>".

Use when the user asks to research, investigate, analyze, find out, explore, examine, audit, or evaluate something — structured multi-source research with consulting, library-doc search, web search, and guided output delivery.

Full autonomous SDLC — design, plan, implement, review, and ship with 4 human approval gates

Drives full multi-agent implementation from a blueprint — fans out parallel-safe tasks into waves, runs sequential tasks in order, resolves conflicts, then verifies. Use to execute an approved plan.

Creates a git worktree for isolated feature work (and documents cleanup post-merge); respects the disableGit preference. Use to isolate a feature on its own branch/worktree.

Multi-agent code review AND remediation — dispatches correctness, security, and simplification reviewers in parallel, consolidates findings, then fixes criticals/highs. Use when reviewing a diff or PR before shipping.

Ships a feature — simplify pass, certify, commit, PR, CI monitoring, review loop, merge. Use when a feature branch is verified and ready to ship.

Security scan — OWASP Top 10, credential/secret detection, injection surfaces, dependency audit, git-history secret scan, auth spot-check. Read-only; produces a severity-ranked report. Use to audit a codebase for vulnerabilities.

Explain Magician status-line availability in Codex. The shipped renderer is Claude Code-specific and this Codex adapter is intentionally a no-op.

Comprehend an existing feature, then PORT it to another app (optionally upgrading it) or INTEGRATE/transform it in place — including swapping the vendor behind the scenes while preserving the exact UX. Use for "port/re-implement/clone this feature into <app>", "swap/migrate the vendor behind <feature> but keep the UX", "figure out how this works then rebuild it", or "walk this flow and recommend improvements".

Systematic debugging with a mandatory hypothesis preflight — no code changes before evidence; one change at a time, then a regression test. Use for any bug, test failure, or unexpected behavior.

TDD engine and enforcer — red/green/refactor, one behavior at a time. Use while implementing any feature or bugfix, or to execute a specific blueprint task with TDD.

Compose and run a large delivery as ONE pipeline with magician's guardrails — for big multi-item work ("implement these N stories/tickets", "deliver the epic", "migrate X across the codebase", batch/sweep). Adaptive structure, but always TDD per unit, kg grounding, certify, multi-lens review + adversarial verify, write gates, no context loss. Use instead of hand-rolling many agents.

Systematic performance profiling and optimization with a mandatory baseline-first gate — measure before changing, re-measure after. Use when something is slow or you need to hit a latency/throughput target.

One-time Codex workspace setup — creates .workspace/ structure, .gitignore entries, a lean AGENTS.md, and suggests relevant Codex capabilities. Run once per project.

Blameless post-mortem / RCA — gathers facts, reconstructs a timeline, runs 5-Whys, defines action items, and writes a post-mortem for review. Use after an incident or outage.

Converts an approved spec into a TDD task plan with a parallelism map (PARALLEL vs SEQUENTIAL), saved to .workspace/shared/plans/. Use after a spec is approved, before implementation.

Full verification loop — tests, types, lint, build, and a Playwright browser check for UI projects; collects evidence before any success claim. Use to verify a change is actually green.

Memory & context steward — session-learning history, the global reference store (repos/projects/ideas), AND live context management (size status, post-compaction resume capsule, project learnings, promotion). Use to review past sessions, remember/recall/forget a reference, check context size, resume after compaction, or capture/consolidate learnings.

Work with Confluence over its REST API (no MCP) — read/search (CQL), summarize pages, find docs, create/update/comment/label. Use for any read/search/create/update on Confluence pages, or references to a remembered space, page, or doc.